Crash Course

This crash course will give you a quick overview of MXNet. You will review core concepts like NDArray (manipulating multiple dimensional arrays) and Gluon (create and train neural networks on CPU and GPU). The intended audience for this crash course is people already familiar with deep learning theory or other deep learning frameworks. For a deep dive into MXNet and deep learning architectures, please refer to [Dive Into Deep learning](http://d2l.ai/) textbook or [Introduction to Deep Learning Course](https://courses.d2l.ai/berkeley-stat-157/index.html)

The course is structured in different sections that can be studied independently or as a whole. If you have a particular question you can consult only the section related to your question, but if you are new to the framework and have time, you can do the course from start to end.

Contents

• Introduction
• Step 1: Manipulate data with NP on MXNet
• Step 2: Create a neural network
• Step 3: Automatic differentiation with autograd
• Step 4: Necessary components that are not in the network
• Step 5: Datasets and DataLoader
• Using own data with included Datasets
• Using your own data with custom Datasets
• New in MXNet 2.0: faster C++ backend dataloaders
• Step 6: Train a Neural Network
• Step 7: Load and Run a NN using GPU

---

Did this page help you?

Yes

No

Thanks for your feedback!

---